Add option to ignore certain dates in parse_date

PAPERLESS_IGNORE_DATES allows to specify a comma separated list of dates
to ignore during date parsing (from filename and content). This can be
used so specify dates that do appear often in documents but are usually
not the documents creation date (like your date of birth).

src/documents/parsers.py

@@ -209,6 +209,13 @@ def parse_date(filename, text):
             }
         )
 
+    def __filter(date):
+        if date and date.year > 1900 and \
+                date <= timezone.now() and \
+                date.date() not in settings.IGNORE_DATES:
+            return date
+        return None
+
     date = None
 
     # if filename date parsing is enabled, search there first:
@@ -222,7 +229,8 @@ def parse_date(filename, text):
                 # Skip all matches that do not parse to a proper date
                 continue
 
-            if date and date.year > 1900 and date <= timezone.now():
+            date = __filter(date)
+            if date is not None:
                 return date
 
     # Iterate through all regex matches in text and try to parse the date
@@ -235,10 +243,9 @@ def parse_date(filename, text):
             # Skip all matches that do not parse to a proper date
             continue
 
-        if date and date.year > 1900 and date <= timezone.now():
+        date = __filter(date)
+        if date is not None:
             break
-        else:
-            date = None
 
     return date
 

src/documents/tests/test_date_parsing.py

@@ -138,3 +138,18 @@ class TestDate(TestCase):
     @override_settings(FILENAME_DATE_ORDER="YMD")
     def test_filename_date_parse_invalid(self, *args):
         self.assertIsNone(parse_date("/tmp/20 408000l 2475 - test.pdf", "No date in here"))
+
+    @override_settings(IGNORE_DATES=(datetime.date(2019, 11, 3), datetime.date(2020, 1, 17)))
+    def test_ignored_dates(self, *args):
+        text = (
+            "lorem ipsum 110319, 20200117 and lorem 13.02.2018 lorem "
+            "ipsum"
+        )
+        date = parse_date("", text)
+        self.assertEqual(
+            date,
+            datetime.datetime(
+                2018, 2, 13, 0, 0,
+                tzinfo=tz.gettz(settings.TIME_ZONE)
+            )
+        )

src/paperless/settings.py

@@ -4,6 +4,7 @@ import multiprocessing
 import os
 import re
 
+import dateparser
 from dotenv import load_dotenv
 
 from django.utils.translation import gettext_lazy as _
@@ -444,3 +445,10 @@ PAPERLESS_TIKA_ENDPOINT = os.getenv("PAPERLESS_TIKA_ENDPOINT", "http://localhost
 PAPERLESS_TIKA_GOTENBERG_ENDPOINT = os.getenv(
     "PAPERLESS_TIKA_GOTENBERG_ENDPOINT", "http://localhost:3000"
 )
+
+# List dates that should be ignored when trying to parse date from document text
+IGNORE_DATES = set()
+for s in os.getenv("PAPERLESS_IGNORE_DATES", "").split(","):
+    d = dateparser.parse(s)
+    if d:
+        IGNORE_DATES.add(d.date())